EMRS DTC Technical Programme Highlights

Document Sample
EMRS DTC Technical Programme Highlights Powered By Docstoc

           Technical Programme Highlights
                       Brian Wardrop
                      Research Director

12/07/05              Pan-DTC Conference               1
What is Required of Remote Sensing?

  The Military Commander needs
     a multi-dimensional (spatial, temporal, spectral) model of his battlespace with
     every object of military relevance accurately located and characterised, and
     an indication of the level of threat it poses to the success of the operation.

  If he attacks a target, he needs
      to know if his attack has hit the target, and
      the functional damage caused to the target, as well as
      any collateral damage caused in the area surrounding the target

  Remote Sensing is required to provide
     Wide area surveillance
     Emitter detection, location, and classification
     Target detection
     Target location and motion
     Target classification and recognition
     Battle damage assessment

        At stand-off ranges sufficient to minimise risk to the sensors

 12/07/05                                Pan-DTC Conference                            2
Scope of EMRS-DTC

 The EMRS-DTC funds:
    Research applicable to military and security systems
    that gather and process electro-magnetic signals,
    propagating in free space to a collecting aperture,
    for the purposes of remote sensing.

 It does not fund research into:
 • Acoustic systems
 • Communication systems (except for IFF).
 • Jamming systems.
 • Directed-energy damage weapons.
 • Systems for detection of chemical & biological agents
 • Data fusion algorithms

 12/07/05                            Pan-DTC Conference    3
Key Regions of the Electromagnetic Spectrum

         103         102         101         100         10-1          10-2      10-3       10-4      10-5       10-6             10-7          10-8          10-9          10-10

Size of a       Football                                    Mouse             Flea             Cell          Bacteria                           Protein          Water
wavelength       Field                                                                                                                                          molecule

Common                                                                                                  Infra-                       Ultra-
                 Radio frequencies                         Microwaves
Name                                                                                                     red                         violet

                                   FM         Mobile Boat
Sources        AM Radio                                                                        People              Light                                               X-ray
                                  Radio       phones radars
                                                                                                                   bulb                                               machines

               106         107         108         109          1010      1011       1012      1013      1014       1015                 1016          1017          1018      1019

                                                   RF                                                        EO
RF Systems                                                                                  EO Systems
   Relatively unaffected by the atmosphere                                                         Affected by scattering and absorption of
   Moving targets can be detected using Doppler shift                                              atmospheric particles, and by atmospheric
   Beamwidths, typically a degree or more, limit                                                   turbulence
   angular resolution.                                                                             High angular resolution
   Passive (listening) systems used to locate and                                                  Surveillance is normally passive using energy
   characterise emitters in key parts of the spectrum                                              reflected or emitted from objects.
                                                                                                   Laser illumination can be used to help track and
                                                                                                   identify targets of interest.
 12/07/05                                                                 Pan-DTC Conference                                                                                          4
EMRS-DTC Programme Themes

                                               Transduction Devices
                                              & Materials (8 Projects)

              RF Systems (16 Projects)

               Transducer Embedded
                                                    EO Systems
              Processing (8 Projects)
                                                    (17 Projects)

 12/07/05                      Pan-DTC Conference                        5
 RF Systems
Science Providers
 University College    Technical Areas
                        Target detection              Military Drivers
 University of
                        Non co-operative target       Target identification
 BAE Systems ATC        recognition
                                                      Detection of new and
 Sula Systems
                                                      difficult target classes
 Birmingham             Networked radar
 University                                           Battle damage
                        Radar and ESM convergence     assessment
 ESL Defence
 Roke Manor Research                                  Improved exploitation
                        Critical system components
                                                      of existing sensor
 TW Research
 RMCS Cranfield

   12/07/05                      Pan-DTC Conference                           6
Cranfield University: Target Detection and Tracking

  The radar signal reflected from air targets of interest are reducing due to:
        The physical size of the target reducing – for example cruise missiles and unmanned air
        The use of geometry and materials when building the aircraft to minimise radar reflections

  From look to look the radar will see a fluctuating target return due to:
        Reflections from different parts of the target reinforcing or cancelling as the target moves
        For targets flying low, reflections from the ground reinforcing or cancelling as the target

 12/07/05                                    Pan-DTC Conference                                        7
Setting the Threshold for Detection

    Detection thresholds must be reduced to
    improve the probability of the target returns
    crossing it, but then false crossings due to noise
    and ground returns confuse the situation,
    making it difficult for tracking algorithms to
    initiate and maintain a track on the target.

    Ground returns and hence false crossings vary
    across the radar cover, requiring the detection
    threshold to be set according to local conditions.                                                         Plot with weak target threshold


                                                                                   Detection threshold for strong targets
                                       Power in range cell


                                                             3                       Detection threshold for weak targets



                                                                 1   5   9   13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97
                                                                                                               Range cell

 12/07/05                                                    Pan-DTC Conference                                                                  8
Heirarchical Approach to Thresholding

 12/07/05                      Pan-DTC Conference   9
BAE SYSTEMS ATC: Non-Cooperative Target Recognition

  Radars are good at detecting potential targets at long
  range in all weathers, but also need to be able to
  classify and identify a target.
  The signal from a target is generated as the
  electromagnetic wave traverses the target and is
  scattered from changes in the target –cockpit, wing
  edges, engines, tail, turbines, etc.
  By using a very short pulse, the signal from each
  scatterer can be isolated, and a ‘range profile’ of the
  target produced. There is, therefore the possibility of
  using the range profile to classify and identify the
  target by comparing it to a library of profiles for
  different targets.
  Since the radar can’t separate returns at the same
  range within the radar beam, the range profile can vary
  dramatically as the aspect of the target changes, and
  so a library of profiles for different aspects is required   Video showing simulation of
                                                               short electromagnetic pulse
  for each target type.                                          illuminating Boeing 707
                                                                  (Courtesy of BAE SYTEMS ATC)

 12/07/05                               Pan-DTC Conference                                       10
Spatial Location of Target Scattering Centres

  Based on the assumption that:
        Target has a number of strong, discrete scattering centres
        These scattering centres are relatively insensitive to
        changes of aspect angle
        The relative positions of the scattering centres can be
        derived from material such as photographs, plans and
        schematics, rather than field measurements.

  Observe target of interest with high range resolution
  waveform over period sufficient to obtain profiles over
  tens of degrees of aspect angle.

  Transform range-angle plot into 2D plan view.

  Locate main scattering centres and compare to library

 12/07/05                                  Pan-DTC Conference        11
 Electro-Optic Systems
Science Providers
 BAE SYSTEMS ATC          Technical Areas
 Thales Optronics
 Thales Optics             Active Imaging
                                                           Military Drivers
 QinetiQ                                                   Enhanced sensing by
 InSys                     Hyperspectral Sensing           using novel concepts
 CST Global                                                and technologies to
 Intense Photonics         Detector Technologies           exploit the electro-
 Waterfall Solutions                                       optical environment.
                           Novel Concepts for EO Sensing
 Institute of Photonics                                    Increased sensor
 St Andrews University                                     performance and
 Heriot-Watt University                                    cost-effectiveness
 Sheffield University                                      through improved
                                                           sensor technologies.
 Imperial College
   12/07/05                         Pan-DTC Conference
BAE SYSTEMS ATC: Optical Power Distribution

  It is anticipated that increasing use will be made of laser
  sources on military aircraft to help in illuminating distant

  Aircraft, and particularly fast jets, have limited space and
  prime power to accommodate a multitude of laser

  The possibility of having a single central laser source
  which can be time-shared between multiple illuminators                      → Near IR PCF
  distributed around the aircraft is, therefore, attractive.                  → Mid IR PCF

  Illuminators often need to be tunable in frequency, and
  this can be achieved by using a device known as a              → Multi-Functional Laser

  parametric oscillator which converts a fixed frequency,        → Optical Parametric Oscillators

  high power input to a lower power, tuneable output. The        → Active EO sensors
  case of interest to this project is where the laser
  produces a high power optical beam which can be routed
  around the aircraft using optical fibres.
 12/07/05                               Pan-DTC Conference                                     13
Modes in Optical Fibres

  Optical fibres made of a simple thread of
  glass can't be used for this application.

  The reason is that when the optical beam is
  coupled into the fibre it can generate and
  propagate as a number of different modes.

  These are characterised by having different
  distributions of power across the cross-
  section of the fibre.

  These modes propagate at different speeds
  down the fibre, and so can produce an
  output beam which varies with temperature
  and flexing of the fibre - unsuitable as an              Cross-section showing power distributions
                                                                       of different modes
  input to the parametric oscillator.

 12/07/05                             Pan-DTC Conference                                               14
Holey Fibres

                                                                                                    Low dielectric
  The higher order modes can be suppressed by using a fibre                                         constant glass
  which has a core made of a doped glass with a higher                                                cladding
  dielectric constant.

  Unfortunately, the optical power, instead of being
  distributed across the fibre, is now concentrated in this tiny
  core, resulting in unacceptable heating when high powers                                              High dielectric
                                                                                                        constant glass
  are transmitted, as well as driving the doped core material                                               core
  into non-linear regions.
                                                                   Cross-section of single mode fibre

  The project is seeking to overcome these limitations by
  using new types of optical fibre which have a cross section
  formed of a regular pattern of holes.
      The spacing of the holes is small compared to the
      optical wavelength, and so to the light, the holey area
      around the core seems to have a different dielectric to
      the rest of the fibre, which is suitable for single mode
      Unlike convention fibres, however, the glass is not
      doped, and can transmit high powers without becoming
      non-linear.                                                        Cross-section of holey fibre

 12/07/05                                   Pan-DTC Conference                                                     15
University of Strathclyde: Adaptive Laser Cavities

•    The performance of active illumination systems using high power solid-state
     lasers depends on the quality and pointing direction of the output beam being
     stable with time.

•    When a laser is switched on, or is subjected to changes in its ambient conditions,
     the output beam can change its characteristics, wandering in angle, and
     changing in intensity.

•    The laser characteristics are determined by the resonant cavity containing the
     lasing material.

•    By incorporating a deformable mirror within the cavity, and continually adjusting it
     to maintain the desired beam characteristics, a more stable source should result.
                                                      Membrane mirror

                             Output coupler                         Nd:GdVO4 slab

                                                              Pump beam

    12/07/05                                  Pan-DTC Conference                            16
 The Deformable Mirror

                         Experimental arrangement

                                                                      Video showing output beam power distribution
                                                                         as system seems optimum mirror shape
                                                                                (Courtesy of University of Strathclyde)

Front face of deformable mirror   Actuators at rear of mirror
    12/07/05                                           Pan-DTC Conference                                                 17
Transduction Devices & Materials
Science Providers

   AREVA                  Technical Areas
   Element6                                              Military Drivers
                           RF Devices and components
                                                         Access to enhanced
   Glasgow University      Multi-layer RF packaging      RF performance at
   Leeds University                                      affordable cost
   QinetiQ                                               provided by compound
                                                         semiconductor devices
   Sheffield University
   UMIST                                                 To improve system
                                                         performance through
   VTT (Finland)
                                                         innovative RF
                                                         components and

   12/07/05                         Pan-DTC Conference                      18
QinetiQ: Wide Bandgap Semiconductors

    Military systems require reliable solid-state microwave power sources.

                   Air SurveillanceRadar
                                                                          Disposable Decoys

                                               Satellite powers amps

                                                                              Mobile Satcomm base stations


                                                       Helicopter ESM

 12/07/05                                  Pan-DTC Conference                                            19
Desirable Material Characteristics

  The performance of power transistors is
  dependent on the properties of the
  semiconducting material used:
        Wide bandgap ? operation at higher
        High breakdown field ? operation at higher
        High thermal conductivity ? operation at
        higher powers
        High mobility and saturated velocitiy ?
        operation at higher frequencies

                                             Si    GaAs        GaN
                   Bandgap (ev)             1.12   1.43        3.4

            Breakdown Field (!06 V/cm)      0.3    0.4          5                                    GaN
        Thermal Conductivity (W/cm/ºK)      1.5    0.5         1.3
        Saturated Velocity (107 cm/sec)      1      1          2.5                                         InP

            Electron Mobility (cm2/V/sec)   1350   8900       1800

 12/07/05                                                 Pan-DTC Conference                                      20
Fabrication Challenges

  Gallium Nitride is grown by depositing it as a thin layer on a disk of material – the
  substrate - inside an evaporating chamber.

  Two factors affect the quality of the grown film:
        Mismatch between the lattice spacing of the Gallium Nitride and that of the substrate
        Mismatch between the thermal expansion of the Gallium Nitride and that of the substrate.

  Silicon substrates are cheap and would reduce manufacturing costs if techniques can
  be developed which minimise the problems of lattice mismatch and thermal expansion.

                                                                       Use of buffer layers to reduce
                  Effects of lattice mismatch
                                                                        effects of lattice mismatch
 12/07/05                                       Pan-DTC Conference                                      21
Transducer Embedded Processing
Science Providers
  Nalletch                                           Technical Areas
  Blue Horizon (GEDAE)                                Rapid development methods
                                                                                     Military Drivers
  Roke Manor Research                                                                To extract maximum
  Thales                                              System Dynamic Range           performance from a
                                                     Enhancement                     sensor before the data is
                                                                                     provided to a bigger
                                                      Signal Processing Algorithm    system
                                                     Development                     Reduce development
                                                                                     cost by reducing design
                                                      Computer Vision                cycle times
                                                                                     To improve design
                                                      Neural Networks                longevity via high level
         Multi-Gigabit                     Megabit
                                                                                     design abstraction
High Speed                                                                           To improve system
Digital Signal                                                                       performance through
                                  Data Processing
 Processing       Multi-Megabit
                                                                                     innovative signal
                                                                                     processing methods
       12/07/05                                                 Pan-DTC Conference
Roke Manor Research: Autonomous Vehicle Navigation

  There is increasing interest in the use
  of autonomous ground vehicles to
  support combat troops:
        As ‘mules’ to carry heavy equipment
        To re-supply ground forces
        To patrol perimeters
        Forward surveillance and mapping

  How will they navigate and plot a
  course through mixed terrain?

  One possibility is to use on-board
  cameras to observe the terrain as the
  vehicle moves, and construct from it a
                                                                 Unmanned test vehicles used by US Army
  3D model which is then used to plot a
                                                               Collaborative Technology Alliance on Robotics
  route.                                                       (photo from

 12/07/05                                 Pan-DTC Conference                                                               23
Structure from Motion

   The Process
      Feature extraction
      Feature tracking
      Object recognition
      Pose determination
      Surface representation
      Drivable regions

 12/07/05                      Pan-DTC Conference   24
Navigation Display

     Autonomous vehicle on test track

                                                                   Detected features colour coded with range

             Clear paths in Cyan

                                                Obstructed paths in Magenta

 12/07/05                               Pan-DTC Conference                                                     25
Video Clip from Laboratory Testing

                             Test vehicle and navigation video
                                 (courtesy of Roke manor Research)

 12/07/05                        Pan-DTC Conference                  26

            Defence Technology Centre


12/07/05        Pan-DTC Conference             27
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

1     EMRS-DTC Technical Programme Highlights

1.1     My presentation to you this afternoon has two objectives.

1.2     The first is to explain the rationale and scope of the Electromagnetic Remote
        Sensing Defence Technology Centre.

1.3     The second is to introduce you to the four technical Themes that make up our
        research programme, and to the projects that we are showcasing to you today
        in the poster rooms.

2     What is Required of Remote Sensing?

2.1     A commander in charge of a military operation needs good situation
        awareness, often in the form of a cognitive model of his battle-space. In this
        model, he would like to have every object of military relevance – buildings,
        vehicles, people – accurately located and characterised, with an estimate of the
        potential threat posed to the success of the operation.

2.2     If a target in the battle-space is attacked, the commander needs to know if it
        has been hit, what functional damage has been inflicted on the target, and any
        collateral damage caused in the area surrounding the target. This information
        will help him to decide if further attacks on the target are needed.

2.3     To help create and maintain this model, as well as provide targeting information
        for weapons systems, remote sensing of the battle-space has a number of
        tasks to undertake, which I have listed at the bottom.

2.4     Since continuity of information flow is key to keeping the commander’s model
        up to date, sensor platform vulnerability is an important factor.

3     Scope of EMRS-DTC

3.1     The scope of the EMRS-DTC is set out here in broad terms, together with
        those areas which we feel are not appropriate for our funding.

4     Key Regions of the Electromagnetic Spectrum

4.1     I have shown here the electromagnetic spectrum that society currently exploits,
        spanning fifteen orders of magnitude in frequency.

4.2     I have indicated those parts of it of particular interest for military remote
        sensing. The Radio Frequency sensors span a range of frequencies from 10
        MHz to 100 GHz – about four orders of magnitude, whilst EO sensors, which
        are normally characterised in terms of wavelength rather than frequency, span
        wavelengths of 100 microns – that is one tenth of a millimetre - to one tenth of
        a micron – about three orders of magnitude.

4.3     Sensors operating in the RF region are relatively unaffected by the
        atmosphere, and make use of the Doppler shift to separate moving targets

Brian Wardrop                                                            20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

        from stationary background. The angular beam-width of RF systems, typically a
        degree or more, does limit angular resolution.

4.4     Since the majority of free-space communication systems use the RF portion of
        the spectrum, an important class of RF military systems are those concerned
        with detecting, locating and exploiting these signals.

4.5     Sensors operating in the EO region, having much shorter wavelengths, are
        affected by scattering and absorption from atmospheric particles, as well as
        atmospheric turbulence. They generally have superior angular resolution to RF
        sensors, and many are passive in operation, detecting light emitted or
        scattered from the scene, although unlike active systems, range is not directly
        available to them.

4.6     If extended operating range, or accurate target ranging is required, this can be
        achieved using an active system, with a laser source illuminating the area of

4.7     It is perhaps worth noting that there is a range of frequencies between RF and
          EO, known as the TeraHertz Gap which at present is little used. However a
          number of UK companies are developing systems in this region which exploit
          its unique molecular interactions with terahertz electromagnetic waves for
          material characterisation, medical diagnostics, process control, and short
          range security systems.

5     EMRS-DTC Programme Themes

5.1     As Dr Wilby has explained in his talk this morning, the portfolio of projects
        funded through the EMRS-DTC are allocated to the four Themes shown here.
        The RF and EO systems Themes are concerned with sensor systems
        operating in these regions, whilst the Transduction Devices and Materials
        Theme is currently concerned with research into RF power devices and
        packaging. The Transducer Embedded Processing Theme is addressing signal
        processing techniques aimed at improving RF and EO sensor performance.

5.2     I will now look at each of these Themes in turn.

6     RF Systems

6.1     This slide shows on the left the organisations currently involved in research
        projects within the Theme.

6.2     Some of the major research areas are shown in the centre, and on the right are
        the military capabilities that we are seeking to satisfy.

6.3     In the EMRS-DTC poster room we have two RF projects on display – one on
        the detection and tracking of small targets, and the other on target recognition.

7     Cranfield University: Target Detection and Tracking

7.1     Radar systems rely on detecting the signal reflected from targets. As the
        physical size of targets of interest decreases – for example cruise missiles and

Brian Wardrop                                                             20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

        unmanned air vehicles – and aircraft design aims to minimise radar reflections
        by shaping and the use of low reflection materials, the returned signal is
        increasingly difficult to detect.

7.2     To complicate the situation, the signal incident on the radar receiver will
        fluctuate with time for reasons shown in the diagram. Signals from different
        parts of the aircraft will interfere sometimes constructively, sometimes

7.3     If the target is close to the ground, signals will also be reflected from the ground
        and similarly re-inforce or cancel with the direct signal.

8     Setting the Threshold for Detection

8.1     In the radar receiver, targets are detected by comparing the received signal
        with a threshold, and a possible detection is declared if the threshold is
        crossed. A sequence of these threshold crossing is used to establish a track,
        and it is only when a track is formed that the presence of a target is accepted.

8.2     Since the signal will be fluctuating for reasons just explained, for some radar
        pulses the target return will not cross the threshold, and this has to be taken
        into account in the tracking.

8.3     The detection of smaller targets and improved tracking can be achieved by
        reducing the threshold. However, as well as the signal from the target, there
        will be thermal noise generated by the radar receiver, as well as signals from
        terrain in the vicinity of the target which are not perfectly suppressed by the
        Doppler processing. Since these ground returns will vary across the radar
        cover, the threshold will need to be varied across the radar cover.

8.4     The plot on the right shows the detections when a low threshold is used – the
        area shown is covered. Several strong returns can be seen forming straight
        tracks, but there is also a weak target which can be made out as a faint pattern
        between the two strong tracks.

9     Heirarchical Approach to Thresholding

9.1     Cranfield University are exploring the use of software agents for detection and
        tracking in situations where low detection thresholds are needed.

9.2     The processing, which is shown schematically on the left has a number of
        levels, taking in the returns from the radar range-azimuth cells shown on the

9.3     Each range-azimuth cell has a level 1 agent which gathers statistics about the
        signal in the cell. Level 2 agents cluster level 1 agents that have similar
        statistics, and use this information to form two thresholds – one high and one
        low - which are used by each level 1 agent.

9.4     If the signal in a range-azimuth cell exceeds the higher threshold, they are
        passed to the conventional radar tracking as likely targets.

Brian Wardrop                                                                20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

9.5     If the signal exceeds the lower threshold, then a detection is declared and a
        level 3 agent created which then strives to form links with existing level 3
        agents. These links represent embryonic tracks.

9.6     The primary function of the level 4 agents is to assess the most likely tracks
        through the level 3 agents, and report the tracks to the main track database if
        they appears to be a true targets.

9.7     This approach not only enables effective use to be made of the available
        computing power using the inherent parallelism of multiple software agents, but
        the self-adaptive, self-organising nature of the agent system should maintain
        performance in an environment with diverse clutter and target characteristics.

10 BAE SYSTEMS ATC: Non-Cooperative Target Recognition

10.1 The second RF poster display is concerned with non-cooperative target
     recognition. The term ‘non-cooperative’ is to distinguish it from civil and military
     air traffic control systems where the aircraft carries a beacon which can be
     interrogated remotely and provides information concerning the aircraft.

10.2 Radars can be very good at detecting targets at long range, but having
     detected a target, some means of classification and identification is needed in
     order to decide if any action is required.

10.3 The majority of the signal reflected from an aircraft is generated from only a few
     points on the aircraft . I have a short video here which shows a very short radar
     pulse – shown as a vertical white bar, and much shorter than the dimensions of
     the aircraft - traversing a Boeing 707. You will see the electromagnetic signal
     being strongly reflected from the noseand the engines, with weaker reflections
     from other parts of the aircraft. It will play twice, the second slowed in time.

10.4 By using a radar that generates such as short pulse, the signal from each
     scatterer can be isolated and a range profile of the target produced - rather like
     a silhouette - which can be compared with a library of profiles to try and identify
     the type of aircraft.

10.5 A major problem with this approach is that the range profile is very sensitive to
     changes in target orientation to the radar, and so a large number of reference
     profiles are required for each target.

11 Spatial Location of Target Scattering Centres

11.1 In an alternative approach, the Advanced Technology Centre of BAE SYTEMS
     is looking at the possibility of processing the radar returns to locate the relative
     positions of the main scattering centres of the target and compare this
     geometrical information to a library of aircraft scattering centres.

11.2 Their approach is based on observing the aircraft with a high range resolution
     waveform over a period of time sufficient for the aspect angle of the aircraft to
     change by tens of degrees. If the returns are plotted as a function of aspect
     angle, the scattering centres are seen as arcs, and the range-angle plot can be
     transformed onto a two-dimensional plan view showing the likely location of the

Brian Wardrop                                                             20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

        main scattering centres. The plots shown here were from a computer model of
        an aircraft, the faint arcs in the 2D plot are generated by the transformation
        algorithm used.

11.3 The approach is based on a number of assumptions: that the target does have
     a number of discrete scattering centres; that they do not change significantly
     with aspect angle, and that the library of scattering centres can be derived from
     material such as photographs and schematics of aircraft of interest. But if
     successful it will provide the commander with much needed information about
     potential threats.

12 Electro-Optic Systems

12.1 Turning now to the Electro-Optic Theme, again we have on the left the
     organisations currently involved in the Theme research.

12.2 Research is addressing four main areas:
•    active imaging, where the target is illuminated by a laser;
•    hyperspectral sensing, where the scene is imaged at a number of separate frequencies
     and the spectral information used to detect and identify targets
•    research aimed at improving the performance and reducing the cost of detectors
•    and novel concepts which fall outside the other three technical areas.

13 BAE SYSTEMS: Optical Power Distribution

13.1 The use of multiple high power infra-red sources on aircraft for target
     illumination and self-defence is expected to increase.

13.2 However aircraft, and particularly fast jets, have difficulty in accommodating a
     multitude of high power lasers due to limitations in space and prime power.

13.3 One option, being explored by the Advanced Technology centre of BAE
     SYTEMS, is to have a single high power laser source which is time-shared
     between systems distributed around the aircraft, the distribution being by an
     optical fibre.

13.4 This is shown in the schematic on the right, where the central laser is shown
     feeding optical parametric oscillators which convert the high power fixed input
     frequency to a lower power tuneable output frequency if needed.

14 Modes in Optical Fibres

14.1 The conventional optical fibre, a fine thread of glass, is unsuitable for this
     application because, when optical power is coupled into its end, the fibre allows
     many different modes to propagate along it. Each mode is characterised by its
     power distribution across the fibre, and some modes are shown here. At the
     bottom left is the fundamental mode, together with some of the higher order
     modes which can also be generated.

14.2 Unfortunately these modes propagate along the fibre with different speeds, and
     as a result at the output of the fibre is a very complex output beam which

Brian Wardrop                                                                 20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

        changes with temperature and flexing of the cable, and is unsuitable for feeding
        a parametric oscillator.

15 Holey Fibres

15.1 It is possible to prevent all but the fundamental mode from propagating by
     using a fibre which has a central core of higher dielectric constant than the
     surrounding glass.

15.2 However this central core, which has an area typically less than 1% of the fibre
     area now carries all of the optical power. Such high power densities can drive
     the doped core material into its non-linear regime.

15.3 The Optoelectronics Research Centre at Southampton University have been
     fabricating a different type of optical fibre known as a holey fibre, the centre of
     which is made up of a matrix of holes surrounding the core.

15.4 Since the core is un-doped, and larger than in a doped single mode fibre, there
     is the possibility that this new type of fibre will be well suited to distributing
     single mode, high power optical signals. The DTC-funded research is
     investigating how to couple high power optical signals into such fibres and the
     propagation characteristics of such fibres when bent into tight loops and when
     vibrated, to see how they would perform in an aircraft environment.

16 University of Strathclyde: Adaptive Laser Cavities

16.1 The robustness and flexibility of solid-state lasers makes them a natural choice
     for active illumination systems.

16.2 In order to provide optical gain, the atoms in the lasing material must be raised
     to higher energy levels, and this is often done by pumping the crystal with a
     high power optical source.

16.3 In many military applications, the laser is switched on only for short periods of

16.4 Unfortunately, when a laser is switched on, it takes time to reach thermal
     equilibrium, and during this time the output beam can change shape and
     pointing direction as the optical path in the resonant cavity changes due to
     refractive index changes through the bulk of the lasing material.

16.5 The University of Strathclyde is looking at the effectiveness of incorporating a
     deformable mirror within the laser cavity to see if the beam characteristics can
     be stabilised against such changes, providing a more stable source.

17 The Deformable Mirror

17.1 The front and back faces of a deformable mirror are shown here at the bottom.
     This is a commercially available device which has had additional mirror
     processing to increase its reflectivity. The actuators, shown to the left, use
     electrostatic attraction to deflect the mirror, each actuator being separately

Brian Wardrop                                                            20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

17.2 A simplified experimental arrangement to explore beam shape adaptation is
     shown at the top. The beam cross section is monitored by software which
     seeks to adjust the actuator voltages to reach an optimum profile.

17.3 The video sequence on the right shows a typical optimisation run, with the
     beam starting quite smeared.

17.4 It is hoped that by incorporating such adaptive elements in the laser cavity, a
     new generation of lasers will result which do not need precision manufacture,
     and are tolerant to environmental changes and ageing of components.

18 Transduction Devices & Materials

18.1 The aim of the Transduction Devices and Materials Theme is to improve
     system performance through innovative RF components and packaging. One
     area of particular interest is microwave power generation.

19 QinetiQ: Wide Bandgap Semiconductors

19.1 QinetiQ are exploring the fabrication of high performance microwave power
     transistors using gallium nitride, which is a wide band-gap semiconductor

19.2 This work has potential application to a range of military systems, some of
     them shown here, which require reliable solid-state sources of microwave

19.3 Current power transistors are fabricated either from silicon - which function best
     at lower frequencies - or gallium arsenide for higher frequency devices.

19.4 As higher powers and higher operating frequencies are sought, there is an
     interest in transistors fabricated from other materials which have more
     desirable characteristics – those that enable operation at high voltages, high
     transistor temperatures, and high frequencies.

20 Desirable Material Characteristics

20.1 Some of the material characteristics that influence the suitability for high power
     operation are shown here.

20.2 One of the most fundamental properties of a material is its band-gap. As shown
     on the upper right, this is the energy required to lift an electron from the
     valence band, to the conduction band. A wide band-gap enables the fabrication
     of transistors which can operate at high temperatures.

20.3 Other parameters such as the field at which the material breaks down, the ease
     with which heat can flow through the material, and the ease with which
     electrons in the conduction band can flow through the material impact transistor

Brian Wardrop                                                           20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

20.4 In the table at the bottom, key material parameters are listed for silicon, gallium
     arsenide, and gallium nitride. Bigger numbers are better, and as can be seen,
     gallium nitride scores high on three of the five.

20.5 The figure at the bottom right is a schematic indication of where gallium nitride
     sits on the power-frequency plane compared with the other compound
     semiconductors such as gallium arsenide, silicon carbide and indium

21 Fabrication Challenges

21.1 It is very difficult to make bulk gallium nitride single crystals in which to
     fabricate transistors, but there is an alternative, which is to deposit a thin layer
     of gallium nitride on a thin disk of a substrate material in an evaporating
     chamber, and then fabricate the transistors in this thin film.

21.2 The performance of the transistors so fabricated are, however, affected by any
     imperfections in the crystal lattice of the gallium nitride.

21.3 The substrate, being of a different material, will differ from the gallium nitride in
     two key aspects. The first is that the lattice spacing of the two materials will be
     different, and the second is that as the substrate is heated, the two materials
     will expand at a different rate.

21.4 The diagram at the bottom shows two cases of lattice match between the
     substrate and deposited film. On the right the two have a matched lattice, whilst
     on the left there is a mismatch resulting in dislocations in the film.

21.5 When fabricating transistors in the film, the substrate must be heated to high
     temperatures and cooled several times, and due to the mismatch in thermal
     expansion rates strains are introduced in the material which can result in
     microcracks in the gallium nitride.

21.6 The objective of the QinetiQ work is to see if device quality gallium nitride film,
     free of dislocations and cracks, can be grown on the six inch silicon wafers
     commonly used by the semiconductor industry. Silicon has a large difference in
     lattice mismatch and thermal expansion, and so complex buffer layers are
     needed, as shown on the right, to ease the mismatch problem.

21.7 The recipe for the buffer layers is complex to work out, but if successful will
     lead to low cost gallium nitride devices for defence.

22 Transducer Embedded Processing

22.1 This Theme is concerned with improving system performance through
     innovative signal processing methods applied to the sensor output signals.

23 Roke Manor Research: Autonomous Vehicle Navigation

23.1 There is growing interest in the use of autonomous ground vehicles to support
     ground troops in combat zones, and some of the suggested roles are shown

Brian Wardrop                                                              20 November 2004
Notes for presentation to Pan-DTC Conference by EMRS-DTC Research Director
Major headings correspond to slide titles

23.2 The picture on the right shows two test vehicles currently being used by the US
     Army in its robotics programme to explore some of these.

23.3 A major challenge for autonomous land vehicles is how to enable them to
     navigate and plot a course through mixed terrain ‘on the fly’, and this is the
     subject of a project by Roke Manor Research.

23.4 Roke are exploring the use of on-board cameras to observe the terrain as the
     vehicle moves, and construct from these images a three dimensional model of
     the surroundings which is then used to plot a route.

24 Structure from Motion

24.1 The generation of structure from motion may be likened to the processing that
     our brain does if we walk around with one eye closed – we locate individual
     objects in our field of view and track their relative motion as we move, enabling
     us to establish their relative distances.

24.2 In the software used here, an image from the camera is processed to isolate
     features – primarily edges and corners – which are tracked form frame to frame
     and used to locate objects and hence driveable regions.

25 Navigation Display

25.1 Before I show you a short film taken during laboratory testing of the software,
     let us have a look at the display generated by the software.

25.2 This is the autonomous vehicle on its test track, which has obstacles,
     boundaries and reference points.

25.3 The display is a perspective view from the vehicle, displaying detected features
     and navigation information. The detected features are colour codes with range,
     blue being far and red near. The navigation information is shown a possible
     paths colour coded according to whether the path is obstructed or clear.

26 Video Clip from Laboratory Testing

26.1 This short sequence shows the vehicle navigating its way to avoid obstacles.

27 END

27.1 I hope that this brief introduction to six of the fifty projects currently supported
     by the EMRS-DTC has given you an indication of the range of work
     undertaken, and will encourage you to talk to the researchers manning their
     poster displays.

Brian Wardrop                                                             20 November 2004